Remotely
datapsychometricsblueprintsrubricsirtcefrvalidity
Also available on
Job Description
📋 Description
- Own Assessment Design — Define what Speak measures across three types; focus on Proficiency Test
- Define Constructs & Build Rubric/Blueprint Layer — Turn goals into concrete constructs and
- Own Validity & the Quality Bar — Sign off on content validity; define mastery criteria; audit
- Design Validity Evidence Plan — Benchmark assessments against external proficiency measures.
- Partner with Product & ML — Collaborate on scoring, calibration, and feedback; balance
🎯 Requirements
- Assessment/Psychometric Design: 4+ years designing rubrics and item specs for language assessments.
- Language Proficiency Domain Expertise: Deep CEFR/ACTFL familiarity.
- Fairness Across Learner Populations: Identify bias and differential item functioning.
- Translate Qualitative to Technical: Convert constructs for ML pipelines.
- Quantitative Rigor: Comfort with reliability, validity, IRT basics.
- Ownership of Quality Bar: Able to sign off on content validity.
🎁 Benefits
- Collaborative environment with ML Engineers and Product teams.
- Global in nature with offices in SF, Ljubljana, Seoul, Tokyo, Taipei.
- Opportunity to shape Speak’s efficacy and impact learners globally.